Update app.py

app.py

@@ -16,12 +16,15 @@ app = FastAPI()
 
 # Use GPU if available
 device = "cuda" if torch.cuda.is_available() else "cpu"
+print(f"✅ Using device: {device}")
 
 # Function to load models lazily
 def get_qa_pipeline():
+    print("🔄 Loading QA pipeline model...")
     return pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", device=device, torch_dtype=torch.float16)
 
 def get_image_captioning_pipeline():
+    print("🔄 Loading Image Captioning model...")
     return pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
 
 ALLOWED_EXTENSIONS = {"pdf", "docx", "pptx", "xlsx"}
@@ -31,44 +34,66 @@ MAX_INPUT_LENGTH = 1024  # Limit input length for faster processing
 def validate_file_type(file):
     if hasattr(file, "name"):
         ext = file.name.split(".")[-1].lower()
+        print(f"📁 File extension detected: {ext}")
         if ext not in ALLOWED_EXTENSIONS:
+            print(f"❌ Unsupported file format: {ext}")
             return f"❌ Unsupported file format: {ext}"
         return None
+    print("❌ Invalid file format!")
     return "❌ Invalid file format!"
 
 # ✅ Extract Text from PDF
 async def extract_text_from_pdf(file):
+    print(f"📄 Extracting text from PDF: {file.name}")
     loop = asyncio.get_event_loop()
-    return await loop.run_in_executor(None, lambda: "\n".join([page.get_text() for page in fitz.open(file.name)]))
+    text = await loop.run_in_executor(None, lambda: "\n".join([page.get_text() for page in fitz.open(file.name)]))
+    print(f"✅ Extracted {len(text)} characters from PDF")
+    return text
 
 # ✅ Extract Text from DOCX
 async def extract_text_from_docx(file):
+    print(f"📄 Extracting text from DOCX: {file.name}")
     loop = asyncio.get_event_loop()
-    return await loop.run_in_executor(None, lambda: "\n".join([p.text for p in Document(file).paragraphs]))
+    text = await loop.run_in_executor(None, lambda: "\n".join([p.text for p in Document(file).paragraphs]))
+    print(f"✅ Extracted {len(text)} characters from DOCX")
+    return text
 
 # ✅ Extract Text from PPTX
 async def extract_text_from_pptx(file):
+    print(f"📄 Extracting text from PPTX: {file.name}")
     loop = asyncio.get_event_loop()
-    return await loop.run_in_executor(None, lambda: "\n".join([shape.text for slide in Presentation(file).slides for shape in slide.shapes if hasattr(shape, "text")]))
+    text = await loop.run_in_executor(None, lambda: "\n".join([shape.text for slide in Presentation(file).slides for shape in slide.shapes if hasattr(shape, "text")]))
+    print(f"✅ Extracted {len(text)} characters from PPTX")
+    return text
 
 # ✅ Extract Text from Excel
 async def extract_text_from_excel(file):
+    print(f"📄 Extracting text from Excel: {file.name}")
     loop = asyncio.get_event_loop()
-    return await loop.run_in_executor(None, lambda: "\n".join([" ".join(str(cell) for cell in row if cell) for sheet in load_workbook(file.name, data_only=True).worksheets for row in sheet.iter_rows(values_only=True)]))
+    text = await loop.run_in_executor(None, lambda: "\n".join([" ".join(str(cell) for cell in row if cell) for sheet in load_workbook(file.name, data_only=True).worksheets for row in sheet.iter_rows(values_only=True)]))
+    print(f"✅ Extracted {len(text)} characters from Excel")
+    return text
 
 # ✅ Truncate Long Text
 def truncate_text(text):
+    print(f"✂️ Truncating text to {MAX_INPUT_LENGTH} characters (if needed)...")
     return text[:MAX_INPUT_LENGTH] if len(text) > MAX_INPUT_LENGTH else text
 
 # ✅ Answer Questions from Image or Document
 async def answer_question(file, question: str):
+    print(f"❓ Question received: {question}")
+
     if isinstance(file, np.ndarray):  # Image Processing
+        print("🖼️ Processing image for captioning...")
         image = Image.fromarray(file)
         image_captioning = get_image_captioning_pipeline()
         caption = image_captioning(image)[0]['generated_text']
+        print(f"📝 Generated caption: {caption}")
 
         qa = get_qa_pipeline()
+        print("🤖 Running QA model...")
         response = qa(f"Question: {question}\nContext: {caption}")
+        print(f"✅ Model response: {response[0]['generated_text']}")
         return response[0]["generated_text"]
 
     validation_error = validate_file_type(file)
@@ -87,18 +112,22 @@ async def answer_question(file, question: str):
     elif file_ext == "xlsx":
         text = await extract_text_from_excel(file)
     else:
+        print("❌ Unsupported file format!")
         return "❌ Unsupported file format!"
 
     if not text:
+        print("⚠️ No text extracted from the document.")
         return "⚠️ No text extracted from the document."
 
     truncated_text = truncate_text(text)
     
     # Run QA model asynchronously
+    print("🤖 Running QA model...")
     loop = asyncio.get_event_loop()
     qa = get_qa_pipeline()
     response = await loop.run_in_executor(None, qa, f"Question: {question}\nContext: {truncated_text}")
 
+    print(f"✅ Model response: {response[0]['generated_text']}")
     return response[0]["generated_text"]
 
 # ✅ Gradio Interface (Separate File & Image Inputs)
@@ -115,6 +144,7 @@ with gr.Blocks() as demo:
     
     submit_btn.click(answer_question, inputs=[file_input, question_input], outputs=answer_output)
 
+
 # ✅ Mount Gradio with FastAPI
 app = gr.mount_gradio_app(app, demo, path="/")